Data driven customer loyalty prediction system and method

Abstract

Systems and methods are provided to predict customer behavior during a digital transaction at a point of sale. The disclosed systems and methods can collect information regarding a merchant and the merchant's business as well as information about the current sales environment in which the merchant is operating. From the collected information, the disclosed systems and methods can process the collected information to generate a prediction of future customer behavior in real-time or near real-time.

What is claimed is: 
     
       1. A method for predicting customer behavior at a point of sale by a payment card processor, the method comprising:
 developing, by the payment card processor, a customer behavior prediction tool by,
 collecting environmental data relating to behavior of a plurality of customers and historical data that includes at least one from among discrete transaction behaviors and purchasing trends of the plurality of customers, the environmental data including a moving sixty-day average transaction amount for the plurality of customers, geographic location data, event data in a calendar for the geographic location, demographic data of the geographic location, social media data, and merchant promotional data; 
 weighting the environmental data based on at least one factor, the at least one factor including a day of the week; 
 automatically generating at least one model based on the weighted environmental data and the historical data by using logistic regression that includes at least one machine learning algorithm; 
 automatically training the at least one model by using a training dataset that includes the weighted environmental data and the historical data; and 
 testing the at least one model by using the training dataset; 
 
 receiving, by the payment card processor, current transaction data related to a digital transaction that includes a cashless transaction between a merchant and a customer at the point of sale; 
 authorizing, by the payment card processor, the digital transaction based on the current transaction data; 
 receiving, by the payment card processor in real-time with the authorizing, historical transaction data related to past purchasing behavior of the customer; 
 identifying, by the payment card processor in real-time with the authorizing, the customer as a desirable customer by,
 determining a correlated data set by using the current transaction data, the historical transaction data, and the at least one model, the correlated data set relating to a correlated change between a first past transaction and a second past transaction; 
 determining a frequency of visits, a spending change during each of the visits, and an average spending amount for the visits by using the correlated data set; 
 generating a threshold curve based on the correlated data set, the frequency of visits, the spending change during each of the visits, and the average spending amount for the visits, the threshold curve relating to an indication of desirability; 
 testing the threshold curve by using a machine learning logistic regression algorithm and the historical transaction data; and 
 applying the machine learning logistic regression algorithm to the current transaction data in real-time; 
 
 displaying, by the payment card processor in real-time via a text message delivered to a mobile device associated with the customer, an approval of the digital transaction based on a result of the authorizing concurrently with at least one reward that includes an immediate benefit for selection based on a result of the identifying; and 
 updating, by the payment card processor, the at least one model by using the current transaction data and the historical transaction data. 
 
     
     
       2. The method of  claim 1 , wherein the current transaction data is received in real-time. 
     
     
       3. The method of  claim 1 , further comprising:
 receiving environmental data and merchant profile data related to the merchant. 
 
     
     
       4. The method of  claim 3 , further comprising:
 generating at least one promotional offer for selection by the merchant to present to the customer at the point of sale. 
 
     
     
       5. The method of  claim 4 , wherein the at least one promotional offer is based on the historical transaction data. 
     
     
       6. The method of  claim 4 , wherein the at least one promotional offer is based on the environmental data related to the merchant. 
     
     
       7. The method of  claim 1 , further comprising:
 generating at least one future purchase prediction of the customer based on the current transaction data and the historical transaction data. 
 
     
     
       8. The method of  claim 7 , wherein identifying the customer as desirable is based on the at least one future purchase prediction. 
     
     
       9. A system for predicting customer behavior at a point of sale by a payment card processor, the system comprising:
 a data processor of the payment card processor; 
 a memory; and 
 a communication interface coupled to each of the data processor and the memory, 
 wherein the data processor of the payment card processor is configured to:
 collect environmental data relating to behavior of a plurality of customers and historical data that includes at least one from among discrete transaction behaviors and purchasing trends of the plurality of customers, the environmental data including a moving sixty-day average transaction amount for the plurality of customers, geographic location data, event data in a calendar for the geographic location, demographic data of the geographic location, social media data, and merchant promotional data; 
 weight the environmental data based on at least one factor, the at least one factor including a day of the week; 
 develop at least one model based on the weighted environmental data and the historical data by using logistic regression that includes at least one machine learning algorithm; 
 train the at least one model by using a training dataset that includes the weighted environmental data and the historical data; 
 test the at least one model by using the training dataset; 
 collect current transaction data related to a digital transaction that includes a cashless transaction between a merchant and a customer at the point of sale; 
 authorize the digital transaction based on the current transaction data; 
 receive historical transaction data related to past purchasing behavior of a customer; 
 identify the customer as a desirable customer by causing the data processor of the payment card processor to:
 determine a correlated data set by using the current transaction data, the historical transaction data, and the at least one model, the correlated data set relating to a correlated change between a first past transaction and a second past transaction; 
 determine a frequency of visits, a spending change during each of the visits, and an average spending amount for the visits by using the correlated data set; 
 generate a threshold curve based on the correlated data set, the frequency of visits, the spending change during each of the visits, and the average spending amount for the visits, the threshold curve relating to an indication of desirability; 
 test the threshold curve by using a machine learning logistic regression algorithm and the historical transaction data; and 
 apply the machine learning logistic regression algorithm to the current transaction data in real-time; 
 
 display, in real-time via a text message delivered to a mobile device associated with the customer, an approval of the digital transaction based on a result of the authorizing together with at least one reward that includes an immediate benefit for selection based on a result of the identifying; and 
 update the at least one model by using the current transaction data and the historical transaction data. 
 
 
     
     
       10. The system of  claim 9 , wherein the current transaction data is received in real time. 
     
     
       11. The system of  claim 9 , wherein the data processor is further configured to collect environmental data and merchant profile data related to the merchant. 
     
     
       12. The system of  claim 11 , wherein the data processor is further configured to generate at least one promotional offer for selection by the merchant to present to the customer at the point of sale. 
     
     
       13. The system of  claim 12 , wherein the at least one promotional offer is based on the environmental data related to the merchant. 
     
     
       14. The system of  claim 12 , wherein the at least one promotional offer is based on the historical transaction data. 
     
     
       15. The system of  claim 14 , wherein the data processor is further configured to generate at least one future purchase prediction of the customer based on the current transaction data and the historical transaction data. 
     
     
       16. The system of  claim 15 , wherein the customer is identified as a desirable customer based on the at least one future purchase prediction. 
     
     
       17. The system of  claim 16 , wherein the data processor is further configured to receive transaction data from an encrypted data cache. 
     
     
       18. The system of  claim 17 , wherein the data processor is configured to decrypt the data received from the encrypted data cache. 
     
     
       19. A method for offering a promotional award by a payment card processor at the point of sale between a customer and a merchant, the method comprising:
 developing, by the payment card processor, a customer behavior prediction tool by,
 collecting environmental data relating to behavior of a plurality of customers and historical data that includes at least one from among discrete transaction behaviors and purchasing trends of the plurality of customers, the environmental data including a moving sixty-day average transaction amount for the plurality of customers, geographic location data, event data in a calendar for the geographic location, demographic data of the geographic location, social media data, and merchant promotional data; 
 weighting the environmental data based on at least one factor, the at least one factor including a day of the week; 
 automatically generating at least one model based on the weighted environmental data and the historical data by using logistic regression that includes at least one machine learning algorithm; 
 automatically training the at least one model by using a training dataset that includes the weighted environmental data and the historical data; and 
 testing the at least one model by using the training dataset; 
 
 receiving, by the payment card processor, current transaction data related to a digital transaction that includes a cashless transaction between the customer and the merchant; 
 authorizing, by the payment card processor, the digital transaction based on the current transaction data; 
 receiving, by the payment card processor in real-time with the authorizing, historical transaction data related to past purchasing behavior of the customer; 
 identifying, by the payment card processor in real-time with the authorizing, the customer as a desirable customer by,
 determining a correlated data set by using the current transaction data, the historical transaction data, and the at least one model, the correlated data set relating to a correlated change between a first past transaction and a second past transaction; 
 determining a frequency of visits, a spending change during each of the visits, and an average spending amount for the visits by using the correlated data set; 
 generating a threshold curve based on the correlated data set, the frequency of visits, the spending change during each of the visits, and the average spending amount for the visits, the threshold curve relating to an indication of desirability; 
 testing the threshold curve by using a machine learning logistic regression algorithm and the historical transaction data; and 
 applying the machine learning logistic regression algorithm to the current transaction data in real-time; 
 
 generating at least one promotional offer for the customer; 
 displaying, by the payment card processor in real-time via a text message delivered to a mobile device, an approval of the digital transaction based on a result of the authorizing concurrently with the at least one promotional offer that includes an immediate benefit for selection based on a result of the identifying; and 
 updating, by the payment card processor, the at least one model by using the current transaction data and the historical transaction data. 
 
     
     
       20. The method of  claim 19 , wherein the current transaction data is received from an encrypted data cache.